Class CRandomFieldGridMap3D

Nested Relationships

Nested Types

Inheritance Relationships

Base Types

  • public mrpt::containers::CDynamicGrid3D< TRandomFieldVoxel >

  • public mrpt::serialization::CSerializable

  • public mrpt::system::COutputLogger

Class Documentation

class CRandomFieldGridMap3D : public mrpt::containers::CDynamicGrid3D<TRandomFieldVoxel>, public mrpt::serialization::CSerializable, public mrpt::system::COutputLogger

CRandomFieldGridMap3D represents a 3D regular grid where each voxel is associated one real-valued property which is to be estimated by this class.

This class implements a Gaussian Markov Random Field (GMRF) estimator, with each voxel being connected to its 6 immediate neighbors (Up, down, left, right, front, back).

  • See papers:

    • ”Time-variant gas distribution mapping with obstacle information”, Monroy, J. G., Blanco, J. L., & Gonzalez-Jimenez, J. Autonomous Robots, 40(1), 1-16, 2016.

Note that this class does not derive from mrpt::maps::CMetricMap since the estimated values do not have sensor-especific semantics, i.e. the grid can be used to estimate temperature, gas concentration, etc.

Usage:

  • Define grid size with either constructor or via setSize().

  • Initialize the map with initialize(). This resets the contents of the map, so previously-added observations will be lost.

  • Add observations of 3D voxels with insertIndividualReading()

Custom connectivity patterns can be defined with setVoxelsConnectivity().

See also

mrpt::maps::CRandomFieldGridMap3D

Public Types

enum TVoxelInterpolationMethod

Values:

enumerator gimNearest
enumerator gimBilinear

Public Functions

CRandomFieldGridMap3D(double x_min = -2, double x_max = 2, double y_min = -2, double y_max = 2, double z_min = -2, double z_max = 2, double voxel_size = 0.5, bool call_initialize_now = true)

Constructor. If you set call_initialize_now to false, the object will be initialized immediately (without the heavy initialization of the GMRF), but you then must call setSize() or clear() later to properly initialize the object before using it to insert observations.

virtual void clear() override

Erases all added observations and start again with an empty gridmap.

bool saveAsCSV(const std::string &filName_mean, const std::string &filName_stddev = std::string()) const

Save the current estimated mean values to a CSV file (compatible with Paraview) with fields x y z mean_value. Optionally, std deviations can be also saved to another file with fields x y z stddev_value, if filName_stddev is provided.

Returns:

false on error writing to file

void resize(double new_x_min, double new_x_max, double new_y_min, double new_y_max, double new_z_min, double new_z_max, const TRandomFieldVoxel &defaultValueNewvoxels, double additionalMarginMeters = 2.0) override

Changes the size of the grid, maintaining previous contents.

See also

setSize

void setSize(const double x_min, const double x_max, const double y_min, const double y_max, const double z_min, const double z_max, const double resolution_xy, const double resolution_z = -1.0, const TRandomFieldVoxel *fill_value = nullptr) override

Changes the size of the grid, erasing previous contents.If resolution_z<0, the same resolution will be used for all dimensions x,y,z as given in resolution_xy

See also

resize.

void setVoxelsConnectivity(const ConnectivityDescriptor::Ptr &new_connectivity_descriptor)

Sets a custom object to define the connectivity between voxels. Must call clear() or setSize() afterwards for the changes to take place.

bool insertIndividualReading(const double sensorReading, const double sensorVariance, const mrpt::math::TPoint3D &point, const TVoxelInterpolationMethod method, const bool update_map)

Direct update of the map with a reading in a given position of the map.

Parameters:

  • sensorReading – [in] The value observed in the (x,y,z) position

  • sensorVariance – [in] The variance of the sensor observation

  • point – [in] The (x,y,z) location

  • method – [in] Voxel interpolation method: how many voxels will be affected by the reading

  • update_map – [in] Run a global map update after inserting this observation (algorithm-dependant)

Returns:

false if point is out of the grid extension.

void updateMapEstimation()

Run the method-specific procedure required to ensure that the mean & variances are up-to-date with all inserted observations, using parameters in insertionOptions

Public Members

TInsertionOptions insertionOptions

See also

updateMapEstimation()

Public Static Attributes

static bool ENABLE_GMRF_PROFILER

[default:false] Enables a profiler to show a performance report at application end.

Protected Functions

void internal_initialize(bool erase_prev_contents = true)

Internal: called called after each change of resolution, size, etc. to build the prior factor information

Protected Attributes

ConnectivityDescriptor::Ptr m_gmrf_connectivity

Empty: default

mrpt::graphs::ScalarFactorGraph m_gmrf
std::vector<std::deque<TObservationGMRF>> m_mrf_factors_activeObs

Vector with the active observations and their respective Information, for each map cell.

std::deque<TPriorFactorGMRF> m_mrf_factors_priors

Vector with the precomputed priors for each GMRF model

struct ConnectivityDescriptor

Base class for user-supplied objects capable of describing voxels connectivity, used to build prior factors of the MRF graph.

See also

setvoxelsConnectivity()

Public Types

using Ptr = std::shared_ptr<ConnectivityDescriptor>

Public Functions

virtual ~ConnectivityDescriptor() = default
virtual bool getEdgeInformation(const CRandomFieldGridMap3D *parent, size_t icx, size_t icy, size_t icz, size_t jcx, size_t jcy, size_t jcz, double &out_edge_information) = 0

Implement the check of whether node i=(icx,icy,icz) is connected with node j=(jcx,jcy,jcy). This visitor method will be called only for immediate neighbors.

Parameters:

  • parent – The parent map on which we are running

  • icx – (cx,cy,cz) for node “i”

  • jcx – (cx,cy,cz) for node “j”

  • out_edge_information – Must output here the inverse of the variance of the constraint edge.

Returns:

true if connected (and the “information” value should be also updated in out_edge_information), false otherwise.

struct TInsertionOptions : public mrpt::config::CLoadableOptions

Parameters common to any derived class. Derived classes should derive a new struct from this one, plus mrpt::config::CLoadableOptions, and call the internal_* methods where appropriate to deal with the variables declared here. Derived classes instantions of their “TInsertionOptions” MUST set the pointer “m_insertOptions_common” upon construction.

Gaussian Markov Random Fields method

double GMRF_lambdaPrior = {0.01f}

The information (Lambda) of fixed map constraints

bool GMRF_skip_variance = {false}

(Default:false) Skip the computation of the variance, just compute the mean

Public Functions

TInsertionOptions()

Default values loader

void loadFromConfigFile(const mrpt::config::CConfigFileBase &source, const std::string &section) override

See mrpt::config::CLoadableOptions

void dumpToTextStream(std::ostream &out) const override

See mrpt::config::CLoadableOptions

struct TObservationGMRF : public mrpt::graphs::ScalarFactorGraph::UnaryFactorVirtualBase

Public Functions

virtual double evaluateResidual() const override
virtual double getInformation() const override
virtual void evalJacobian(double &dr_dx) const override
inline TObservationGMRF(CRandomFieldGridMap3D &parent)

Public Members

double obsValue

Observation value

double Lambda

“Information” of the observation (=inverse of the variance)

struct TPriorFactorGMRF : public mrpt::graphs::ScalarFactorGraph::BinaryFactorVirtualBase

Public Functions

virtual double evaluateResidual() const override
virtual double getInformation() const override
virtual void evalJacobian(double &dr_dx_i, double &dr_dx_j) const override
inline TPriorFactorGMRF(CRandomFieldGridMap3D &parent)

Public Members

double Lambda

“Information” of the observation (=inverse of the variance)